Posted By Kent McMillan on 5/22/2008 at 1:34 AM

 

Warning! Discussion of survey measurements and analysis of uncertainties to follow!

 

 

One of the things that I find to be fairly consistently true of surveying in older, developed districts is how relatively difficult it is to connect to all of the boundary monuments that are needed for the survey. Landscaping, fences, building improvements, other obstacles, and cranky landowners all conspire to make things tougher than they would be in that ideal world that has yet to arrive.

 

In a thread the other day, I posted a sketch of some of the parcels resurveyed.

 

As a professional surveyor, naturally I thought it was important to perpetuate the positions of the boundary markers found and set. This was done by two methods.

 

a) As many traverse and network control points were marked in concrete structures as possible. A control point like Spike and Washer 53 shown in the photo below should be in place and usable for more than 40 years. The spike is a 3/8 in. x 8 in. steel spike driven into a drill hole in the curb and with a punchmark on the center of the head for the station mark.

 

b) Accurate geodetic grid coordinates were determined for all points positioned by the survey.

Here is an image of the control network that ended up getting surveyed on that project:

 

The light blue lines from 57 to 55 and off screen are the GPS vectors. All of the rest was not really GPSable and was surveyed conventionally with a total station.

 

The usual residential hazards of fences etc. meant that three markers adjacent on the parcel boundaries, 140, 122, and 124, (shown on the sketch linked above) ended up being located by methods that would not have been used on a wide open, unrestricted site.

 

One of the tasks in the office was to verify that the relative uncertainties of the three points were not excessive. This was where least squares survey adjustment software with Star*Net's error analysis features served very well. Well it served very well in adjusting the GPS vectors in combination with the conventional survey measurements between the same points. The uncertainty analysis was better yet.

 

What appears to be a closed loop, i.e. 58-54-53-11-51-60-59-112-58, about 1450 ft. in length, wasn't a closed traverse in the conventional sense. To negotiate a fence, control points 60 and 61 inside a 6 ft. privacy fence enclosure were carefully tied from outside it and then the instrument station inside the fence was determined by resection from Point 59 and another fence was negotiated to get a redundant tie to Point 112 on the other side of yet another 6 ft. privacy fence that had been tied directly from Point 58 on the other side of the block.

 

There was no angular closure easily possible, but the coordinate misclosure of the position of 112 coming from one directoin compared to that of 112 from the other direction was part of the means of verifying that no blunders were present.

 

What uncertainty analysis showed was that the differences in 112 were entirely consistent with the predictions based upon the uncertainties in the angles, distances, etc. that were measured. Everything was as it was expected to be.

 

The net result that that check was consistent with was that all points positioned by the survey ended up with NAD 83 (CORS96) Epoch 2002.0 coordinates with standard errors in N and E that were under +/-0.015 ft. It was light work to check the relative uncertainties between Rod 140, Pipe 122, and Pipe 124 to confirm that the inversed bearings and distances were well within tolerance.

 

Best regards,
Kent McMillan, RPLS Austin TX
Modified By Kent McMillan on 5/22/2008 at 2:07